💎 Google unveils Gemini 3.1 Pro

Good Morning! Google just dropped a new Gemini upgrade that boosts performance and undercuts rivals on price, tightening the cost curve in the frontier model race. Plus, you will learn how to use Nano Banana Pro to create 10 unique thumbnail concepts for your next video.

Plus, in today’s AI newsletter:

• Google unveiled Gemini 3.1 Pro

• Rork Max One-Shots Apple Apps

• Perplexity Joins the Anti-Ad Camp

• How to Create 10 Unique Thumbnail Concepts for Your Next Video

• 4 new AI tools worth trying

AI MODELS

💎 Google unveiled Gemini 3.1 Pro

Google unveiled a preview of Gemini 3.1 Pro, positioning it as a new baseline for complex reasoning, multimodal work, and long-horizon tasks, while pricing it below competing frontier models.

• Beats models from Anthropic and OpenAI across reasoning, scientific knowledge, agentic coding, and tool use benchmarks

• Supports text, image, audio, and video inputs with up to a 1M token context window and 64K token outputs

• Costs $2 / 1M input tokens and $12 / 1M output tokens, cheaper than Claude Opus 4.6 and competitive with GPT-5.2

• Available in Gemini app, NotebookLM, and across enterprise tools like Vertex AI, AI Studio, and Android Studio

As AI costs become a bottleneck, Google is betting that cheaper frontier models win long-term. If it keeps pushing performance while compressing prices, Gemini could quietly become the most widely adopted high-end model across both consumer and enterprise stacks.

AI TOOLS

🍎 Rork Max One-Shots Apple Apps

Rork introduces Rork Max. It’s an AI that can build, install, and publish full-fledged apps across the entire Apple ecosystem, without Xcode.

• Builds iPhone, iPad,  Watch,  TV & Vision Pro apps (AR, 3D games, body tracking, Live Activities)

• Replaces Xcode with a web app: 1-click install, 2-click App Store publish
  Can recreate complex apps like Pokémon Go, Minecraft-style worlds, and

• Vision Pro games in hours

• Powered by Swift, Claude Code, Opus 4.6, with Siri, widgets, and on-device Vision AI

Rork isn’t just speeding up development, it’s removing the biggest barrier to building for Apple platforms. If this works at scale, solo creators could ship App Store-level products without ever opening Xcode.

AI NEWS

🧭 Perplexity Joins the Anti-Ad Camp

Perplexity says ads risk making users “doubt everything,” and is now focusing on paid products instead. The shift highlights a growing divide in AI: ads vs accuracy.

• Perplexity has phased out ads and isn’t pursuing new ad deals right now

• Executives say ads conflict with trust and accuracy in AI answers

• The company is doubling down on paid users like enterprises, finance, and legal professionals

• Ads aren’t fully ruled out long-term, but seen as misaligned with user expectations

AI companies are hitting a monetization crossroads. Perplexity is betting that users will pay for unbiased, trustworthy AI, while others lean into ads to fund massive compute costs. How this shakes out could define which AI platforms people actually trust.

HOW TO AI

🗂️ How to Create 10 Unique Thumbnail Concepts for Your Next Video Without a Designer

In this tutorial, you will learn how to use Nano Banana Pro, a high-fidelity image reasoning engine, to generate professionally designed YouTube thumbnails.

🧰 Who is This For

• Content Creators who want to increase their CTR (Click-Through Rate) without hiring a pro.

• Solopreneurs needing high-quality branding for social media.

• Marketers who need to A/B test multiple thumbnail concepts quickly.

• Beginners who struggle with Photoshop or complex design layouts.

STEP 1: Access Nano Banana Pro

Open your Gemini workspace and navigate to the Tools menu. Select “🍌 Create images” and ensure you switch the model toggle to “Nano Banana Pro” (sometimes labeled as the "Thinking" model). You will see a prompt box designed for advanced visual reasoning.

STEP 2: Use the "Identity Lock" Prompt

One of Nano Banana Pro’s superpowers is keeping your face consistent.

• Action: Upload a high-quality headshot of yourself.

• The Prompt: “Use the person from this image. Keep the facial features identical, but change the expression to [Shocked/Excited/Confused]. Place the person on the left third of the frame, pointing toward the right.”

This ensures you look like yourself across all 10 concepts while adding the high-energy emotion needed for a viral thumbnail.

STEP 3: Generate the "Batch 10" Concept

Instead of one image, ask the model to reason through different layouts based on proven YouTube psychology.

You can use this prompt: “Generate 10 unique thumbnail variations for a video about '[Your Topic]'. Use different compositions: one with a 'Before vs After' split, one with a 'Glowing Mystery Box', and one with 'Bold 3D Neon Text' reading '[Your Title]'. Ensure 4K resolution and high-contrast colors.”

Nano Banana Pro will "think" through the spatial requirements and present you with variations that include legible, rendered text.

STEP 4: Refine with "Natural Language Edits"

If you like concept #3 but the background is too busy, don't restart. Use the multi-turn editing feature.

Type directly into the chat: “I like the 3rd one, but change the background to a dark, moody tech studio with purple neon accents. Also, make the yellow arrow larger and more vibrant.”

Once satisfied, click the Download button to save your thumbnail in native 2K or 4K resolution, ready for immediate upload to YouTube.

Meta says it is “explicitly separating” Quest VR from its Worlds platform, making Worlds “almost exclusively mobile”, letting it better compete with Roblox.

Microsoft's AI safety team proposed technical standards for detecting AI-generated content, but its CSO declined to commit to using them across its platforms.

Microsoft confirms a bug that let Microsoft 365 Copilot summarize confidential emails from Sent Items and Drafts folders, and deployed a fix.

Ex-DeepMind researcher David Silver is reportedly raising $1B at a $4B valuation for his London-based AI startup Ineffable Intelligence, in Europe’s largest ever seed round.

🧑‍💻 Gemini 3.1 Pro: Google’s new, cheaper AI for complex and multimodal tasks

🌎 Tiny Aya: Cohere's small, open-source model covering 70+ languages

🤖 Manus: AI agents you can use on Telegram that remember things over time

🎶 Lyria 3: Google’s AI that makes music inside Gemini

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas